A method and device for identifying deformation of concrete dam
Through the combination of the MM estimation model and the minimum covariance determinant, the outlier discrimination accuracy in concrete dam deformation monitoring is improved, the problem of low discrimination accuracy in the prior art is solved, and the monitoring factors and load factors are realized Effective distinction enhances the reliability of structural analysis and maintenance.
Patent Information
- Application Number
- CN202411095205.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-12
AI Technical Summary
In the deformation monitoring of concrete dams, the accuracy of the deformation outliers is low, making it difficult to effectively distinguish outliers caused by monitoring factors and load factors.
The MM-estimation model predicts deformation value, and determines the robust distance based on the minimum covariance determinant of the deformation influence factor. Combining the deformation residual and the stable distance, we can distinguish whether the deformation is an outlier, and then distinguishes the coarse outlier and the load outlier. value.
It improves the accuracy of the identification of deformation outliers, can more accurately distinguish outliers caused by monitoring factors and load factors, enhances the reliability of concrete dam structure analysis and maintenance, and reduces the probability of accidents.
Smart Images

Figure CN118643577B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of deformation monitoring of concrete dams, and in particular to a deformation identification method and device of concrete dams. Background Art
[0002] Usually, concrete dams will deform under the action of water pressure, temperature and other loads. Deformation is the most intuitive reflection of the structural form of concrete dams, so the deformation of concrete dams needs to be monitored.
[0003] Due to the influence of certain factors, the deformation monitoring values of concrete dams may be abnormal, and the abnormal monitoring values are considered to be deformation outliers. At present, in the process of distinguishing deformation outliers in deformation monitoring values, there are cases of misjudgment of deformation outliers, resulting in low accuracy in distinguishing deformation outliers. Summary of the invention
[0004] The present invention provides a deformation identification method and device for a concrete dam, which can improve the accuracy of identifying deformation outliers during the identification of deformation monitoring values of the concrete dam.
[0005] In order to achieve the above object, the present invention adopts the following technical solution:
[0006] In a first aspect, the present invention provides a method for identifying deformation of a concrete dam, comprising: obtaining a deformation monitoring value of the concrete dam;
[0007] The deformation of the concrete dam is predicted by using the MM estimation model to obtain the predicted deformation value of the concrete dam;
[0008] Based on the minimum covariance determinant of the deformation influencing factor of the concrete dam, the robust distance of the deformation influencing factor of the concrete dam is determined;
[0009] When the deformation residual of the concrete dam is less than or equal to the residual threshold, and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold, the deformation monitoring value is determined to be a normal value; otherwise, the deformation monitoring value is determined to be an outlier; wherein the deformation residual of the concrete dam is the absolute value of the difference between the deformation monitoring value and the deformation prediction value.
[0010] The deformation identification method of the concrete dam provided by the present invention determines the deformation prediction value by using the MM estimation method, and determines the robust distance of the deformation influence factor of the concrete dam based on the minimum covariance determinant of the deformation influence factor of the concrete dam; then, the deformation residual of the concrete dam (the absolute value of the difference between the deformation monitoring value and the deformation prediction value of the concrete dam) and the robust distance of the deformation influence factor of the concrete dam are combined to determine whether the deformation is an outlier. Compared with the scheme in the prior art that predicts the deformation value based on the least squares method and identifies the outlier according to the deformation residual, the method provided by the present invention comprehensively considers the deformation residual and the robust distance of the deformation influence factor, and can improve the identification accuracy of the deformation outlier to a certain extent.
[0011] In an implementation manner of the first aspect, when the deformation monitoring value is an outlier, the method includes:
[0012] Determine the categories of the above outliers; the categories of outliers include gross error outliers and load outliers. Gross error outliers are outliers caused by monitoring factors, and load outliers are outliers caused by load factors.
[0013] In an implementation of the first aspect, determining a category of an outlier includes:
[0014] When the deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is greater than the robust distance threshold, the deformation monitoring value is determined to be a gross error outlier;
[0015] When the deformation residual of the concrete dam is greater than the residual threshold and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold, the deformation monitoring value is determined to be a load outlier;
[0016] When the deformation residual of the concrete dam is less than or equal to the residual threshold and the robust distance of the deformation influencing factor is greater than the robust distance threshold, the deformation monitoring value is determined to be a load outlier.
[0017] In an implementation of the first aspect, the deformation residual of the concrete dam obeys a chi-square distribution, and the residual threshold is ,in, The degrees of freedom are k、 The significance level is α Critical value of the chi-square distribution.
[0018] In an implementation of the first aspect, the minimum covariance determinant estimator includes an estimated value of a mean vector of the deformation influencing factor and an estimated value of a covariance matrix of the deformation influencing factor; and the robust distance of the deformation influencing factor of the concrete dam satisfies:
[0019] ;
[0020] in, It is a concrete dam n The first i A measurement value, Deformation influence factor The robust distance is the estimated value of the mean vector of deformation influencing factors; is the estimated value of the covariance matrix of the deformation influencing factors; for The transpose of for The inverse matrix of
[0021] satisfy: ; express The Mahalanobis distance from the center of the deformation influence factor, represents the weight function; , is the mean of the subset of n deformation influencing factor measurements that makes the covariance matrix have the minimum determinant, Represents the covariance matrix and consistency factor with the minimum determinant The product of
[0022] satisfy:
[0023] ; is the coefficient of consistency.
[0024] In an implementation of the first aspect, the robust distance of the deformation influencing factor of the concrete dam obeys a chi-square distribution, and the robust distance threshold is ;in, The degrees of freedom are k、 The significance level is α Critical value of the chi-square distribution.
[0025] In a second aspect, the present invention provides a deformation identification device for a concrete dam, comprising an acquisition module, a prediction module, a determination module and an identification module;
[0026] The acquisition module is used to obtain the deformation monitoring value of the concrete dam;
[0027] The prediction module is used to predict the deformation of the concrete dam using the MM estimation model to obtain the deformation prediction value of the concrete dam;
[0028] The determination module is used to determine the robust distance of the deformation influence factor of the concrete dam based on the minimum covariance determinant of the deformation influence factor of the concrete dam;
[0029] The identification module is used to determine that the deformation monitoring value is a normal value when the deformation residual of the concrete dam is less than or equal to the residual threshold and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold; otherwise, the deformation monitoring value is determined to be an outlier; wherein the deformation residual of the concrete dam is the absolute value of the difference between the deformation monitoring value and the deformation prediction value.
[0030] In an implementation of the second aspect, the identification module is specifically used to determine the category of outliers; the categories of outliers include gross error outliers and load outliers, gross error outliers are outliers caused by monitoring factors, and load outliers are outliers caused by load factors.
[0031] In a third aspect, the present invention provides an electronic device, comprising a processor and a memory coupled to the processor; the memory is used to store computer instructions, and when the electronic device is running, the processor executes the computer instructions stored in the memory, so that the electronic device executes the method as described in the first aspect or any one of its implementations.
[0032] In a fourth aspect, the present invention provides a computer-readable storage medium, comprising computer program instructions, which, when executed by a computer, enable the computer to execute the method in the first aspect or any one of its implementations.
[0033] In a fifth aspect, the present invention provides a computer program product, comprising computer program instructions, which, when executed on a computer, enable the computer to execute the method in the first aspect or any one of its implementations.
[0034] The technical effects corresponding to the above-mentioned second to fifth aspects and their possible implementations can refer to the above-mentioned description of the technical effects of the first aspect and its possible implementations, and will not be repeated here.
[0035] The beneficial effects of the present invention are:
[0036] The deformation identification method of the concrete dam provided by the present invention determines the deformation prediction value by using the MM estimation method, and determines the robust distance of the deformation influence factor of the concrete dam based on the minimum covariance determinant of the deformation influence factor of the concrete dam; then determines whether the deformation is an outlier by combining the deformation residual of the concrete dam with the robust distance of the deformation influence factor of the concrete dam. Compared with the scheme in the prior art that predicts the deformation value based on the least squares method and identifies the outlier according to the deformation residual, the method provided by the present invention comprehensively considers the deformation residual and the robust distance of the deformation influence factor, and can improve the identification accuracy of the deformation outlier to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] Figure 1It is one of the schematic diagrams of the deformation identification method of the concrete dam provided by the embodiment of the present invention;
[0038] Figure 2 This is a second schematic diagram of a deformation identification method for a concrete dam provided by an embodiment of the present invention;
[0039] Figure 3 This is a third schematic diagram of a deformation identification method for a concrete dam provided by an embodiment of the present invention;
[0040] Figure 4 It is a schematic diagram of the operation flow of MM estimation in the deformation identification method of a concrete dam provided by an embodiment of the present invention;
[0041] Figure 5 is a deformation identification diagram in the deformation identification method of a concrete dam provided by an embodiment of the present invention;
[0042] Figure 6 It is a schematic diagram of the association between deformation influencing factors and deformation monitoring values in the deformation identification method of a concrete dam provided by an embodiment of the present invention;
[0043] Figure 7 It is a structural schematic diagram of a deformation identification device for a concrete dam provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0044] The terms "first" and "second" etc. in the description and claims of the present invention are used to distinguish different objects rather than to describe a specific order of the objects.
[0045] “And / or” in the embodiment of the present invention represents the relationship between objects. For example, A and / or B can represent the following three situations: A exists alone, B exists alone, and A and B exist at the same time.
[0046] In the embodiments of the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.
[0047] The method and device provided by the embodiment of the present invention can be used to identify the deformation monitoring value of the concrete dam obtained by monitoring, so as to identify whether the deformation monitoring value is an outlier. Furthermore, the type of the outlier can also be identified (i.e., the factor causing the outlier can be identified).
[0048] During the service of concrete dams, the dam structure may be deformed due to various factors such as temperature changes, foundation settlement, reservoir water level changes, and seismic activities. In order to ensure the normal use of concrete dams, deformation monitoring is required. When monitoring the deformation of concrete dams, some abnormal deformation values are often detected. In order to ensure the accuracy of engineering monitoring and subsequent maintenance of concrete dams, it is necessary to identify abnormal values (i.e. outliers) in the deformation values of concrete dams.
[0049] It is understandable that the deformation outliers of concrete dams may be caused by human errors and monitoring instruments during the monitoring process, such as manual reading or recording errors, improper calibration of sampling instruments, instrument failures, etc. This type of outlier is random and unpredictable, and its essence is the gross error of deformation. The existence of this type of outlier seriously interferes with the quality of deformation behavior analysis and tracking of concrete dams. Therefore, this type of outlier should be eliminated.
[0050] The abnormal deformation value of the concrete dam may also be caused by load factors such as water pressure and temperature, as well as the time factor caused by long-term load. The change law of this type of outliers is crucial to analyzing the deformation behavior of the dam. Based on the outliers caused by these load reasons, the deformation degree of the concrete dam and its possible consequences can be structurally analyzed and predicted, so that the concrete dam can be better maintained and the probability of accidents can be reduced. Therefore, the identification and screening of outliers caused by load changes / sudden changes has high research value and practical significance.
[0051] At present, it is usually difficult to distinguish outliers caused by load changes / sudden changes from outliers caused by human errors / monitoring instruments. Through the embodiments of the present invention, outliers can be distinguished into gross error factors and load factors. In summary, the deformation outliers of concrete dams can include outliers caused by monitoring factors and outliers caused by load factors.
[0052] In the prior art, an identification model based on a statistical model is usually used to identify the deformation value of a concrete dam. The identification model aims to minimize the residual error and uses the least squares estimation method to identify the deformation monitoring value of the concrete dam. This method is to manually select deformation influencing factors. In order to improve the accuracy, the identification model requires enough input data, and the identification model will adapt to all data including outliers as much as possible, which in turn reduces the prediction accuracy of the model for normal values. In other words, the model cannot reflect the impact of outliers and ensure high accuracy. Therefore, it is easy to produce misjudgment (normal values are judged as outliers) and missed judgment (outliers are judged as normal values), resulting in low accuracy in identifying deformation outliers.
[0053] In order to solve the problem of low accuracy in identifying deformation outliers in the background technology, an embodiment of the present invention provides a deformation identification method and device for a concrete dam, which can combine the deformation residual of the concrete dam (the absolute value of the difference between the deformation monitoring value and the deformation prediction value of the concrete dam) and the robust distance of the deformation influencing factor of the concrete dam to determine whether the deformation is an outlier, thereby improving the accuracy of identifying deformation outliers.
[0054] First, the technical terms involved in the deformation identification method and device for a concrete dam provided in an embodiment of the present invention are explained.
[0055] Deformed outlier identification model:
[0056] This model is a deformation outlier identification model based on statistical theory, which can be used to predict the deformation of object structures, and is usually used to predict the structural deformation of dams, bridges, buildings, etc. For example, taking a concrete dam as an example, by establishing the relationship between the deformation of the concrete dam and the factors causing the deformation, an identification model is obtained. Based on the identification model, the deformation value of the concrete dam (hereinafter referred to as the deformation prediction value) can be predicted when the factors causing the deformation are known.
[0057] According to the description of the above embodiment, the deformation of the concrete dam is mainly caused by three factors (load factors): water pressure, temperature and aging. Therefore, the deformation value of the concrete dam (denoted as δ ) includes deformation components caused by three factors, specifically, the deformation component corresponding to the water pressure (hereinafter referred to as the water pressure component, recorded as ), the deformation component corresponding to temperature (hereinafter referred to as temperature component, denoted as ), and the deformation component corresponding to aging (hereinafter referred to as aging component, recorded as ). Then the deformation outlier identification model is:
[0058] ;
[0059] 1. Water pressure component :
[0060] The above-mentioned water pressure components include the deformation of the dam body itself caused by the static water pressure, the deformation of the dam caused by the internal force generated on the dam base surface, and the deformation caused by the rotation of the dam base surface due to the gravity of the reservoir water.
[0061] Through analysis, the deformation of any point in the above concrete dam and reservoir depth (i.e. upstream reservoir water level) related to, specific to, with It is a linear relationship, and the expression of water pressure component is:
[0062] (1);
[0063] in, is the coefficient, is the water level of the upstream reservoir. For different dam types, The values are different. For example, for a gravity dam, It can be taken as 3. For arch dams, 4 or 5 is acceptable.
[0064] It should be understood that the is the deformation influencing factor affecting the water pressure component;
[0065] 2. Temperature component :
[0066] Temperature component The deformation is mainly caused by the temperature change of the dam body concrete and the dam foundation. The temperature change of the dam can be reflected by the temperature measurement values buried in the dam body and the dam foundation. There are generally two situations for the layout of thermometers:
[0067] 1) When there are enough thermometers, their measured values can reflect the temperature changes;
[0068] In one implementation, the temperature component is determined based on the measured values of each thermometer. :
[0069] (2);
[0070] in, For the i Thermometer measurement value, For the i The coefficients of the measured values, is the number of thermometers. is the deformation influencing factor that affects the temperature component.
[0071] In another implementation, the temperature component is determined based on the equivalent temperature (the equivalent temperature is the average value of the thermometer measurement value) :
[0072] (3);
[0073] in, For the i The equivalent temperature of the layer, For the i The gradient change value of the equivalent temperature change of the layer, For the i The coefficient of the equivalent temperature of the layer, For the iThe coefficient of the gradient change value of the equivalent temperature change of the layer, is the number of layers with equivalent temperature.
[0074] In the above formula (3), and is the deformation influencing factor affecting the temperature component;
[0075] 2) In the case of few thermometers (i.e. lack of thermometer measurements), if the heat of concrete hydration has been dissipated and the temperature field inside the dam has reached a relatively stable level, the temperature component can be determined based on multi-period harmonics. :
[0076] (4);
[0077] In the formula, For the cycle, Represents the annual cycle, Indicates a half-year cycle, generally Choose 1 or 2; , is the coefficient; It is the cumulative number of days from the start of measurement to the monitoring day.
[0078] In the above formula (4), and Deformation influence factors affecting the temperature component;
[0079] 3. Timeliness :
[0080] Water pressure component and temperature component It is the elastic deformation component of any point on the concrete dam. Under long-term load, the dam also produces deformation that changes with time, that is, time-dependent deformation. The expression is:
[0081] (5);
[0082] In the formula, , is the coefficient; , It is the cumulative number of days from the start of measurement to the monitoring day.
[0083] In summary, in the first implementation, combining the above formulas (1), (2) and (5), the deformation outlier identification model of the concrete dam is:
[0084] (6);
[0085] In the above formula, is the deformation influencing factor;
[0086] In the second implementation, combining the above formulas (1), (3) and (5), the deformation outlier identification model of the concrete dam is:
[0087] (7);
[0088] In formula (7), is the deformation influencing factor;
[0089] In the third implementation, combining the above formulas (1), (4) and (5), the deformation outlier identification model of the concrete dam is:
[0090] (8);
[0091] In formula (8), is the deformation influencing factor;
[0092] In summary, taking the deformation value of a certain measuring point of the concrete dam as the dependent variable and the deformation influencing factor as the independent variable, the deformation outlier identification model based on statistical theory described in the above formulas (6), (7) and (8) can be rewritten as:
[0093] ;
[0094] in, is the predicted deformation value of the concrete dam, X represents the deformation influencing factor (including multiple deformation influencing factors), represents the error term;
[0095] The above identification model is used to predict the deformation value of the concrete dam. i The deformation prediction value is:
[0096] (9);
[0097] In the formula, For concrete dam i The deformation prediction value, for The i deformation influence factor vector, is the regression coefficient vector, For the i The error term of the deformation prediction value;
[0098] Exemplarily, corresponding to formula (6), the deformation influence factor vector is:
[0099] ;
[0100] The above regression coefficient vector is:
[0101] ;
[0102] Corresponding to formula (7), the deformation influence factor vector is:
[0103] ;
[0104] The above regression coefficient vector is:
[0105] ;
[0106] Corresponding to formula (8), the deformation influence factor vector is:
[0107] ;
[0108] The above regression coefficient vector is:
[0109] .
[0110] Exemplarily, the method for deformation identification of a concrete dam provided in an embodiment of the present invention may be executed by an electronic device having a processing function, for example, the electronic device may be a computer, a server, etc. Taking the electronic device as a computer as an example, the hardware part of the computer may include: a processor, a memory, a network interface, a user interface, a communication bus, etc.
[0111] The processor is used to control the electronic device to perform related processing and computing tasks. For example, in the embodiment of the present invention, the processor is used to predict the deformation of the concrete dam according to the MM estimation model and to identify the deformation monitoring value based on the MM estimation result. The processor may include a central processing unit (CPU). i ng un i t, CPU) or other processors, the processor may be single-core or multi-core, for example, the processor may include multiple CPUs.
[0112] The memory is used to store computer instructions and related data, for example, to store the deformation monitoring value data of the concrete dam in the embodiment of the present invention, and the deformation prediction value data calculated by the processor. The memory can be a random access memory (RAM), a read only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, or an optical memory, a disk storage medium or other magnetic storage device, or any other medium that can be used to store program codes or data that can be accessed by a computer. Optionally, the memory can be integrated into the processor, and the memory can also be independent of the processor.
[0113] The network interface is used for the computer to communicate with other devices or communication networks. The network interface can be a transceiver with transceiver functions. Optionally, the network interface can include a standard wired interface, a wireless interface (such as a WI-FI interface, a Bluetooth interface, a 5G interface).
[0114] The communication bus is used to realize the connection and communication between various components. For example, the processor, memory, network interface and user interface mentioned above can be interconnected through the communication bus.
[0115] The user interface may include a display screen and an input unit (such as a keyboard). Optionally, the user interface may also include a standard wired interface and a wireless interface.
[0116] Those skilled in the art will appreciate that the above-mentioned computer may also include more or fewer components, or a combination of certain components, or different arrangements of components, which is not limited in the embodiments of the present invention.
[0117] In combination with the above, the deformation identification method of a concrete dam provided by an embodiment of the present invention is described in detail below.
[0118] like Figure 1 As shown, the deformation identification method of a concrete dam provided by the embodiment of the present invention includes S101-S104;
[0119] S101, obtaining a deformation monitoring value of a concrete dam;
[0120] It should be understood that the above deformation monitoring values are obtained by monitoring instruments. The staff uses the detection instruments to measure the deformation of the concrete dam and obtain multiple deformation monitoring values of the concrete dam over a period of time.
[0121] Optionally, the deformation monitoring value may be any one of a horizontal displacement monitoring value, a vertical displacement monitoring value, a crack monitoring value, etc.; illustratively, a horizontal deformation monitoring value of a perpendicular line of a concrete dam arch seat along the upstream and downstream directions of the dam may be selected as the deformation monitoring value of the concrete dam;
[0122] S102, using the MM estimation model to predict the deformation of the concrete dam to obtain a predicted deformation value of the concrete dam;
[0123] The MM estimation (Modified Maximum Likelihood Estimator, MM-estimator) method is a method used for parameter estimation in statistics. The method is mainly used to process data sets with outliers or outliers. The MM estimation modifies the maximum likelihood estimation to make it insensitive to the influence of outliers, thereby obtaining a more robust estimation result. In the embodiment of the present invention, the MM estimation is used to predict the deformation monitoring value of the concrete dam to obtain a robust deformation prediction value.
[0124] In the embodiment of the present invention, MM estimation includes S estimation and M estimation, and the regression coefficient in the identification model is obtained by S estimation; the regression coefficient obtained by S estimation is used as the initial value of the regression coefficient, and the deformation prediction value is obtained by M estimation. The deformation prediction value can be used to identify abnormal values in the deformation monitoring value;
[0125] Optional, combined Figure 1 ,like Figure 2 As shown, the above S102 includes S1021-S1022;
[0126] S1021, with the goal of minimizing the discreteness of the deformation residual, the regression coefficient and the deformation residual in the identification model are obtained through S estimation;
[0127] In the embodiment of the present invention, the first of the multiple deformation monitoring values is i The deformation monitoring value is recorded as δ i , and the deformation prediction value is recorded as , deformation monitoring value δ i and deformation prediction value The difference is the deformation residual of the concrete dam (denoted as e i , ), refer to the identification model shown in the above formula (9) ( ), the deformation residual of the concrete dam can be expressed as:
[0128] (10);
[0129] in, For the i deformation influence factor vector, is the regression coefficient vector The predicted value of , the error term in the identification model is ignored;
[0130] The process of S estimation is: n The discreteness of the deformation residuals is minimized as the solution goal, and the predicted value of the regression coefficient vector is calculated and deformation residual e i ;
[0131] The objective function for minimizing the discreteness of n deformation residuals is:
[0132] (11);
[0133] In formula (11), Represents the regression coefficient vector as The i The deformation residual (i.e. ), Express Perform discrete minimization processing;
[0134] Among them, in the specific calculation process of minimizing discreteness, the deformation residual function must meet the following conditions:
[0135] (12);
[0136] In the formula, is the deformation residual function, is a constant, , where Φ represents the standard normal distribution, that is It should obey the standard normal distribution, and the expectation of the standard normal distribution is ;
[0137] Above is a first-order weight function, optional, It can be any of the bisquare weight function, Huber weight function, cauchy weight, etc.
[0138] Under the constraints shown in formula (12), by solving formula (11), we can get the predicted value of the regression coefficient vector , and then get the deformation residual ; Subsequently, the estimated S and As the initial value of the regression coefficient vector and the initial value of the deformation residual in the M estimation process (i.e., the initial value of the M estimation and );
[0139] In one embodiment, the deformation residual function is a Huber weight function:
[0140] (13);
[0141] In the formula, is the fine tuning constant; according to the measurement, take =1.345, the relative efficiency is close to 95%, indicating that the deformation outlier identification model has the best effect in resisting the interference of outliers. Therefore, The value can be 1.345;
[0142] It is understandable that when the deformation residual is in the center of the distribution (i.e. ), the Huber weight function is the minimum sum of squares, and the weight function Give the same weight to water pressure, temperature and aging; when the residual value is distributed at both ends (i.e. ), the weight function is transformed into the minimum absolute value, and the weight assigned to the residuals farther out of the tail of the distribution is smaller. By distinguishing the residual weights mentioned above, the obtained It can more truly reflect the influence of various factors, thereby improving the prediction accuracy of the deformation outlier identification model;
[0143] In another embodiment, the deformation residual function is a bi-square weight function:
[0144] (14);
[0145] In the formula, the fine tuning constant is The value of is 4.685; according to the calculation, when the fine-tuning constant =4.685, it can produce 95% efficiency, indicating that the deformation outlier identification model has the best effect in resisting the interference of outliers at this time;
[0146] S1022, using the regression coefficient vector and deformation residual obtained by S estimation as initial values of the regression coefficient vector and deformation residual, taking the minimum sum of deformation residual function values as the goal, and using M estimation to obtain a deformation prediction value;
[0147] M estimation is an estimation method of robust regression. The principle of M estimation is: construct a residual function and calculate the regression coefficient vector that minimizes the sum of the deformed residual function values; obtain the initial value of the regression coefficient vector used for M estimation through the above S1021 and the initial value of the deformation residual , using M estimation, using the weight function to perform multiple iterations of weighted least squares on the deformation residual, to obtain the converged deformation residual, thereby obtaining the regression coefficient vector, and then substituting the regression coefficient vector into the identification model to obtain the deformation prediction value of the concrete dam ;
[0148] The objective function of M estimation is:
[0149] (15);
[0150] In the formula, is the deformation residual function, is a strictly convex function with a low increasing rate; the solution of this function is It does not have scale equivariance, that is, ,in a >0, X A random variable representing a deformation influencing factor;
[0151] Optionally, the above deformation residual function It can be a Huber weight function or a bi-square weight function, which is not limited in the embodiment of the present invention;
[0152] The deformation residual in the above formula (15) is e i The standardized value of , hereinafter referred to as standardized residuals) should satisfy the following conditions:
[0153] (16);
[0154] In the formula, is the weight function of the standardized residual;
[0155] The calculation formula for the standardized residual is as follows:
[0156] (17);
[0157] In the formula, MAD is the median absolute deviation (Med i an Absolute Dev i at i on), , is the median of multiple deformation monitoring values;
[0158] In the embodiment of the present invention, the weight function of the above-mentioned standardized residual satisfies:
[0159] (18);
[0160] in, is the deformed residual function of the standardized residual The partial derivative of with respect to the regression coefficient is satisfy:
[0161] (19);
[0162] Optionally, when the deformation residual function is a Huber weight function, the weight function of the standardized residual is for:
[0163] (20);
[0164] In the formula, is the fine tuning constant;
[0165] Optionally, when the deformation residual function is a bisquare weight function, the weight function of the standardized residual is for:
[0166] (twenty one);
[0167] In the formula, is the fine tuning constant.
[0168] Based on the above, the solution process of M estimation includes S1-S4:
[0169] S1. Calculate the initial value of the deformation residual The corresponding initial value of the standardized residual ;
[0170] S2. Substituting into the weight function of the standardized residual, we get ;
[0171] The weight function of the standardized residual may be the function shown in the above formula (20) or formula (21);
[0172] S3. Substituting into formula (16), the weighted least squares method is used to solve it and we get , update the standardized residual to , and further obtain the deformation residual , thus obtaining ;
[0173] S4, with the updated standardized residual As input, repeat steps S2-S3 and perform multiple iterative calculations until the regression coefficient vector converges (for example, the convergence condition can be that the residual between the regression coefficient vector obtained in this iteration and the regression coefficient vector obtained in the previous iteration is less than a preset value), and obtain the estimated value of the regression coefficient vector , and then according to Get the deformation prediction value of the concrete dam ;
[0174] Figure 4 The complete flow chart of MM estimation is shown, which can be combined with Figure 4 The above is an understanding of the contents of S1021-S1022;
[0175] In the embodiment of the present invention, the initial values of the regression coefficient vector and the deformation residual are obtained by S estimation, and the iterative weighted least square method is used in M estimation to continuously update the regression coefficient vector, so that the regression coefficient vector converges in the direction of reducing the residual in each iteration step, ensuring that the algorithm eventually converges to the global optimal solution, providing a more accurate and robust regression coefficient estimation, and thus obtaining a deformation prediction value with higher accuracy;
[0176] S103, determining a robust distance of the deformation influencing factor of the concrete dam based on a minimum covariance determinant estimator of the deformation influencing factor of the concrete dam;
[0177] The robust distance in the embodiment of the present invention is a distance measurement used to measure the difference in the change of data points. The larger the robust distance is, the greater the difference between a data point and the data center of multiple data points is, which can indicate that the data point is an outlier data point to a certain extent.
[0178] In the embodiment of the present invention, the robust distance is determined based on the idea of the minimum covariance determinant (MCD) and the Mahalanobis distance. The minimum covariance determinant can be understood as the covariance matrix of the deformation influencing factor having the minimum determinant.
[0179] Specifically, the minimum covariance determinant estimator is first calculated, and the minimum covariance determinant estimator includes an estimated value of the mean vector of the deformation influencing factor and an estimated value of the covariance matrix of the deformation influencing factor (the covariance matrix has a minimum determinant);
[0180] The deformation influencing factor of concrete dam is , , n for The number of measured values, is the deformation influencing factor i A monitoring sequence, pRepresentative i The number of influencing factors in a monitoring sequence;
[0181] Estimated value of the mean vector of deformation influence factors satisfy:
[0182] ;
[0183] express and The Mahalanobis distance between the centers of represents the weight function; , yes n The subset of the deformation influencing factor measurements that makes the covariance matrix have the minimum determinant The mean of Represents the covariance matrix and consistency factor with the minimum determinant The product of
[0184] Estimated value of the covariance matrix of deformation influencing factors satisfy:
[0185] ; is the coefficient of consistency;
[0186] In the embodiment of the present invention, n The measured values of the deformation influencing factors are taken as a data set, and the data set is divided into multiple sub-data sets, each of which includes h ([( n + p + 1) / 2]≤ h ≤ n ); Find the sub-dataset with the smallest determinant of the covariance matrix from multiple sub-datasets, calculate the mean of the sub-dataset, and get ; Then the consistency factor Multiplying with the covariance matrix of the above data set (i.e., the covariance matrix with the minimum determinant) yields ;
[0187] Optionally, the above weight function It can be: , is the indicator function, The degrees of freedom are k、 The significance level is α (e.g. α = 0.05) is the critical value of the chi-square distribution; when Less than or equal to (i.e. the input is true), the value of the weight function is 1; when Greater than (i.e. the input is false), the value of the weight function is 0;
[0188] Based on the above, the robust distance of the deformation influencing factor of the concrete dam satisfies:
[0189] ;
[0190] in, It is a concrete dam n The first i A measurement value, Deformation influence factor The robust distance is the estimated value of the mean vector of deformation influencing factors; is the estimated value of the covariance matrix of the deformation influencing factors; for The transpose of for The inverse matrix of is the coefficient of consistency;
[0191] In one implementation, the C-step (concentrat i on step) selection method, select the sub-dataset with the covariance matrix with the smallest determinant (including h deformation influence factor measurements) in order to calculate the Mahalanobis distance using this sub-dataset , and then use Calculated , ;
[0192] Specifically: n In the data set consisting of the measured values of deformation influencing factors, first randomly select h ( h < n ) deformation influencing factor measurements as sub-datasets H 1 , the number of deformation influence factors in each deformation influence factor measurement (dimension) p Greater than or equal to 2; illustratively, n ≤600, the number of repetitions in the algorithm is set to 500, h =10;
[0193] ①. First calculate h The initial mean vector of the measured values of the deformation influencing factors , initial covariance matrix ;
[0194] Initial mean vector satisfy: ;
[0195] Initial covariance matrix satisfy: ;
[0196] ②If det( )≠0, calculate the measured value of each deformation influencing factor in the data set and Distance:
[0197] ;
[0198] in, i The value is 1 , … ,n ;det( ) represents the covariance matrix The determinant of ;
[0199] ③、Compare the measured value of each deformation influence factor obtained by the above calculation with Distance d 1 ( i ) Sort in ascending order, and then select the first h The smaller deformation influencing factor measurements constitute the sub-dataset H 2 ;based on H 2 Recalculate the mean vector and covariance matrix, that is , , det( )≤ det( );
[0200] Step ② to step ③ in the FAST-MCD algorithm is a C-step;
[0201] ④、Repeat the above steps until det( )=0 or det( )=det( ), the final result is the covariance matrix with the minimum determinant , the mean vector with the minimum determinant covariance matrix ;
[0202] In another implementation, if the data volume is large, the data set can be randomly split into several subsets, and then the above-mentioned FAST-MCD algorithm is executed for each subset, and the mean vector of the entire data set and the covariance matrix with the minimum determinant are obtained according to the processing results of each subset;
[0203] For example,n = 1000 as an example, 5 subsets can be constructed, and the number of deformation influencing factor measurements in each data subset n sub = 200, the above FAST-MCD algorithm is executed for each subset, and the number of repetitions is 100. For each subset, the amount of data of the deformation influencing factor measurements in the constructed sub-data set is h sub (For example, h sub =4);
[0204] Step 1: For each of the five subsets, perform the above C-step twice to obtain ;
[0205] Step 2: Select the first 10 smaller deformation influence factor measurements that make the covariance matrix have the minimum determinant from the 200 deformation influence factor measurements of each data subset, and merge the 10 deformation influence factor measurements extracted from each of the five subsets into one sub-dataset;
[0206] Step 3: For the merged sub-dataset, set h = 10, and the FAST-MCD algorithm is performed on the merged sub-dataset to select 10 deformation influence factor measurements that make the covariance matrix have the minimum determinant from the 50 deformation influence factor measurements, and obtain , .
[0207] In the embodiment of the present invention, the distance calculated based on the minimum covariance determinant estimator (robust mean vector and robust covariance matrix) is insensitive to abnormal values or outliers (i.e., has good robustness to abnormal values), that is, the distance can remain relatively stable and will not fluctuate greatly due to the influence of a few extreme values;
[0208] According to the robust distance of the influencing factors in the deformation prediction value obtained in S102 and the deformation monitoring value obtained in S103, it is determined whether the deformation monitoring value obtained by monitoring is an abnormal value (i.e., a deformation outlier) through the following S104;
[0209] S104. When the deformation residual of the concrete dam is less than or equal to the residual threshold, and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold, the deformation monitoring value is determined to be a normal value; otherwise, the deformation monitoring value is determined to be an outlier; wherein the deformation residual of the concrete dam is the absolute value of the difference between the deformation monitoring value and the deformation prediction value;
[0210] It should be understood that the smaller the deformation residual, the closer the deformation monitoring value is to the deformation prediction value, that is, the higher the credibility of the deformation monitoring value; the smaller the robust distance, the smaller the change difference of the deformation influencing factor;
[0211] In the embodiment of the present invention, the residual threshold can be calculated according to the probability distribution type obeyed by the deformation residual of the concrete dam; the robust distance threshold can be calculated according to the probability distribution type obeyed by the robust distance of the deformation influencing factor;
[0212] In one implementation, the deformation residual of the concrete dam follows a chi-square distribution, and the residual threshold is , The degrees of freedom are k、 The significance level is α The critical value of the chi-square distribution can be obtained by looking up the chi-square distribution table;
[0213] Specifically, the deformation residual of the concrete dam is less than or equal to the residual threshold value, which includes: the difference between the deformation monitoring value and the deformation prediction value is less than or equal to , and the difference between the deformation monitoring value and the deformation prediction value is greater than or equal to ;
[0214] In some implementations, the deformation residual in S102 may be a corrected deformation residual, and the corrected residual is , is the deformation residual before correction, is the deformation prediction value The standard deviation of
[0215] In one implementation, the robust distance of the deformation influencing factor of the concrete dam follows a chi-square distribution, and the robust distance threshold is ;in, The degrees of freedom are k、 The significance level is α The critical value of the chi-square distribution of ;
[0216] Optionally, the probability distribution type may also be other distribution types, such as normal distribution, binomial distribution, or Poisson distribution, which is not limited in the embodiment of the present invention.
[0217] In the embodiment of the present invention, the deformation residual of the concrete dam is less than or equal to the residual threshold, indicating that the difference between the deformation detection value and the deformation prediction value is small, and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold, indicating that the change difference of the deformation influencing factor is small. In this case, it can be determined that the deformation monitoring value is relatively reliable, and the deformation monitoring value is not an outlier caused by the monitoring factor (i.e., gross error) or the load factor, and the deformation monitoring value is a normal value;
[0218] In summary, the deformation identification method of the concrete dam provided by the embodiment of the present invention uses the MM estimation method to determine the deformation prediction value, and determines the robust distance of the deformation influence factor of the concrete dam based on the minimum covariance determinant of the deformation influence factor of the concrete dam; then, the deformation residual of the concrete dam (the absolute value of the difference between the deformation monitoring value and the deformation prediction value of the concrete dam) and the robust distance of the deformation influence factor of the concrete dam are combined to determine whether the deformation is an outlier. Compared with the scheme in the prior art that predicts the deformation value based on the least squares method and identifies the outlier according to the deformation residual, the method provided by the present invention comprehensively considers the deformation residual and the robust distance of the deformation influence factor, and can improve the identification accuracy of the deformation outlier to a certain extent;
[0219] Combination Figure 1 ,like Figure 3 As shown, when the deformation monitoring value determined by S101-S104 is an outlier, the deformation outlier identification method provided by the embodiment of the present invention further includes S105;
[0220] S105, when the deformation monitoring value is an outlier, determining the category of the outlier;
[0221] In combination with the above S104, it can be seen that when the deformation residual and deformation influencing factor of the concrete dam meet one of the following conditions, the deformation monitoring value is an outlier:
[0222] Condition 1: The deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold;
[0223] Condition 2: The deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is greater than the robust distance threshold;
[0224] Condition 3: The deformation residual of the concrete dam is less than the residual threshold, and the robust distance of the deformation influencing factor is greater than the robust distance threshold;
[0225] The categories of outliers include gross error outliers and load outliers. Gross error outliers are outliers caused by monitoring factors, and load outliers are outliers caused by load factors.
[0226] In one implementation, the residual threshold +e (i.e. + )、 -e (Right now- ), robust distance threshold ( ), distinguish between gross error outliers and load outliers;
[0227] Specifically, Figure 5As shown in the figure, the robust distance of the deformation influencing factor is used as the horizontal coordinate and the deformation residual is used as the vertical coordinate to define the outlier identification coordinate system. The deformation residual threshold on the vertical axis of the outlier identification coordinate system is +e and -e , and the robust distance threshold on the horizontal axis can divide the outlier identification coordinate system into 6 regions, which are recorded as region I, region II, region III, region IV, region V, and region VI;
[0228] In region I, the deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold;
[0229] In region II, the deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is greater than the robust distance threshold;
[0230] In region III, the deformation residual of the concrete dam is less than the residual threshold, and the robust distance of the deformation influencing factor is less than the robust distance threshold;
[0231] In region IV, the deformation residual of the concrete dam is less than the residual threshold, and the robust distance of the deformation influencing factor is greater than the robust distance threshold;
[0232] In region V, the deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold;
[0233] In region VI, the deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is greater than the robust distance threshold;
[0234] It should be understood that when the robust distance between the deformation residual and the deformation influencing factor falls into the above region III, the deformation monitoring value is a normal value, and the region III is a normal value area, that is, the deformation residual is small and the robust distance is small; the deformation value falling into region I, region II, region IV, region V, and region VI is an outlier; the regions corresponding to the above condition 1 are regions I and region V, the regions corresponding to condition 2 are regions II and region VI, and the region corresponding to condition 3 is region IV;
[0235] like Figure 6 As shown, the horizontal axis represents the change of deformation influencing factor X, and the vertical axis represents the deformation monitoring value δ changes in Figure 6 The closer the data points in the coordinate system shown are to the dotted line, the smaller the deformation residual is. Figure 5 and Figure 6 , the data points in area III (circular data points) are close to the dotted line, and the independent variable X is not far away from the main data. The data in area III are normal data, that is, the deformation monitoring value is normal;
[0236] In area IV (triangle data points), area II, and area VI (cross star data points), because the independent variable X is far away from the main data, the data points in area IV, area II, and area VI are leverage points;
[0237] Among them, if the data points in region IV are close to the dotted line, then the data points in region IV are good leverage points, and it is determined that the deformation outlier is caused by the load factor, that is, the deformation outlier is a load outlier;
[0238] The data points in regions II and IV are far from the dotted line, and the independent variable X is far from the main data. The data points in regions II and VI are bad leverage points. It is determined that the deformed outlier is caused by the gross error factor, that is, the deformed outlier is a gross error outlier.
[0239] The data points in area I and area V are far away from the dotted line, and the independent variable X is not far away from the main data. Then the data points in area II and area VI are vertical outliers. It is determined that the deformation outlier is caused by the load factor, that is, the deformation outlier is a load outlier.
[0240] In addition, according to the performance differences between the robust distances of deformation residuals and deformation influencing factors caused by gross error factors and load factors, the residual threshold and the robust distance threshold are used to delineate the area, which can further distinguish whether the gross error outliers are load outliers, and facilitate the subsequent screening of deformation monitoring values caused by gross error factors and analysis of deformation monitoring values caused by load factors, thereby facilitating structural analysis and prediction of the status of the concrete dam, and thus better maintaining the concrete dam and reducing the probability of accidents.
[0241] Accordingly, an embodiment of the present invention provides a deformation identification device for a concrete dam, such as Figure 7 As shown, it includes an acquisition module 501, a prediction module 502, a determination module 503 and an identification module 504;
[0242] The acquisition module 501 is used to acquire the deformation monitoring value of the concrete dam; for example, the acquisition module 501 is used to implement S101 in the above-mentioned deformation identification method of the concrete dam;
[0243] The prediction module 502 is used to predict the deformation of the concrete dam using the MM estimation model to obtain a predicted deformation value of the concrete dam; for example, the prediction module 502 is used to implement S102 in the above-mentioned method for identifying the deformation of the concrete dam;
[0244] The determination module 503 is used to determine the robust distance of the deformation influencing factor of the concrete dam based on the minimum covariance determinant of the deformation influencing factor of the concrete dam; for example, the determination module 503 is used to implement S103 in the above-mentioned deformation identification method of the concrete dam;
[0245] The identification module 504 is used to determine that the deformation monitoring value is a normal value when the deformation residual of the concrete dam is less than or equal to the residual threshold and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold; otherwise, the deformation monitoring value is determined to be an outlier; wherein the deformation residual of the concrete dam is the absolute value of the difference between the deformation monitoring value and the deformation prediction value; for example, the identification module 504 is used to implement S104 in the above-mentioned concrete dam deformation identification method.
[0246] Each module of the above-mentioned concrete dam deformation identification device can also be used to execute other steps in the above-mentioned method embodiment. The relevant contents involved in the above-mentioned method embodiment can all be referred to the functional description of the corresponding functional modules, which will not be repeated here.
[0247] The embodiment of the present invention also provides an electronic device, including: a processor and a memory coupled to the processor; the memory is used to store computer instructions, and when the electronic device is running, the processor executes the computer instructions stored in the memory, so that the electronic device executes the method in the above embodiment. Among them, the processor can implement the above acquisition module 501, prediction module 502, determination module 503 and identification module 504; the above memory can also be used to store the monitoring data obtained by the acquisition module 501, the prediction data obtained by the prediction module 502, the robust data obtained by the determination module 503 and the identification result obtained by the identification module 504.
[0248] An embodiment of the present invention further provides a computer-readable storage medium, which includes a computer program. When the computer program runs on a computer, the method described in the above embodiment is executed.
[0249] An embodiment of the present invention further provides a computer program product, which includes computer program instructions. When the computer program instructions are run on a computer, the method described in the above embodiment is executed.
[0250] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
[0251] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for identifying deformation of a concrete dam, characterized in that: include: S101, obtaining a deformation monitoring value of a concrete dam; S102, using the MM estimation model to predict the deformation of the concrete dam to obtain the deformation prediction value of the concrete dam; wherein, using the MM estimation model to predict the deformation of the concrete dam to obtain the deformation prediction value of the concrete dam includes: taking the discreteness of the deformation residual as the goal, obtaining the regression coefficient and the deformation residual in the identification model through S estimation; using the regression coefficient vector and the deformation residual obtained by S estimation as the initial values of the regression coefficient vector and the deformation residual, taking the sum of the deformation residual function value as the goal, and using M estimation to obtain the deformation prediction value; The objective function of minimizing the discreteness of the deformation residual is: In the formula, Represents the regression coefficient vector as The i-th deformation residual (i.e. ), Express Perform discrete minimization to obtain the initial value of the regression coefficient vector and the initial value of the deformation residual The objective function with the minimum sum of deformation residual function values as the goal is: In the formula, ρ(e i ) is the deformation residual function, is the predicted value of the regression coefficient vector β; x i is the i-th deformation influence factor vector, e i is the deformation residual, δ i Deformation monitoring value; S103, determining a robust distance of the deformation influencing factor of the concrete dam based on the minimum covariance determinant of the deformation influencing factor of the concrete dam; wherein the robust distance of the deformation influencing factor of the concrete dam satisfies: Among them, xi is the ith measured value among the n measured values of deformation influencing factors of concrete dam, D R (x i ) is the robust distance of the deformation influence factor xi, T MCD is the estimated value of the mean vector of the deformation influencing factor; C MCD is the estimated value of the covariance matrix of the deformation influencing factors; (x i -T MCD ) T is (x i -T MCD ) is the transpose of MCD -1 C MCD The inverse matrix of T MCD satisfy: d i represents the Mahalanobis distance between xi and the center of the deformation influencing factor, w(*) represents the weight function; is the mean of the subset of n deformation influencing factor measurements that makes the covariance matrix have the minimum determinant, represents the product of the covariance matrix with the minimum determinant and the consistency coefficient c0; C MCD satisfy: c1 is the consistency coefficient; S104, when the deformation residual of the concrete dam is less than or equal to the residual threshold, and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold, determining that the deformation monitoring value is a normal value; otherwise, determining that the deformation monitoring value is an outlier; wherein the deformation residual of the concrete dam is the absolute value of the difference between the deformation monitoring value and the deformation prediction value; When the deformation monitoring value is an outlier, the method includes: Determine the category of the outlier; the category of the outlier includes a gross error outlier and a load outlier, the gross error outlier is an outlier caused by a monitoring factor, and the load outlier is an outlier caused by a load factor; Determining the category of the outlier includes: When the deformation residual of the concrete dam is greater than the residual threshold, and the robust distance of the deformation influencing factor is greater than the robust distance threshold, determining that the deformation monitoring value is a gross error outlier; When the deformation residual of the concrete dam is greater than the residual threshold and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold, determining the deformation monitoring value as a load outlier; When the deformation residual of the concrete dam is less than or equal to the residual threshold and the robust distance of the deformation influencing factor is greater than the robust distance threshold, the deformation monitoring value is determined to be a load outlier.
2. The method according to claim 1, characterized in that The deformation residual of the concrete dam follows a chi-square distribution, and the residual threshold is in, represents the critical value of the chi-square distribution with k degrees of freedom and α as the significance level.
3. The method according to claim 1, characterized in that The robust distance of the deformation influencing factor of the concrete dam follows the chi-square distribution, and the robust distance threshold is in, represents the critical value of the chi-square distribution with k degrees of freedom and α as the significance level.
4. A deformation identification device for a concrete dam, used to implement the deformation identification method for a concrete dam according to any one of claims 1 to 3, characterized in that: It includes an acquisition module, a prediction module, a determination module and an identification module; The acquisition module is used to obtain the deformation monitoring value of the concrete dam; The prediction module is used to predict the deformation of the concrete dam using the MM estimation model to obtain the deformation prediction value of the concrete dam; wherein, the prediction of the deformation of the concrete dam using the MM estimation model to obtain the deformation prediction value of the concrete dam includes: taking the discreteness of the deformation residual as the goal, obtaining the regression coefficient and the deformation residual in the identification model through S estimation; taking the regression coefficient vector and the deformation residual obtained by S estimation as the initial values of the regression coefficient vector and the deformation residual, taking the sum of the deformation residual function value as the goal, and using M estimation to obtain the deformation prediction value; The determination module is used to determine the robust distance of the deformation influence factor of the concrete dam based on the minimum covariance determinant of the deformation influence factor of the concrete dam; wherein the robust distance of the deformation influence factor of the concrete dam satisfies: Among them, xi is the ith measured value among the n measured values of deformation influencing factors of concrete dam, D R (x i ) is the robust distance of the deformation influence factor xi, T MCD is the estimated value of the mean vector of the deformation influencing factor; C MCD is the estimated value of the covariance matrix of the deformation influencing factors; (x i -T MCD ) T is (x i -T MCD ) is the transpose of MCD -1 C MCD The inverse matrix of T MCD satisfy: d i represents the Mahalanobis distance between xi and the center of the deformation influencing factor, w(*) represents the weight function; is the mean of the subset of n deformation influencing factor measurements that makes the covariance matrix have the minimum determinant, represents the product of the covariance matrix with the minimum determinant and the consistency coefficient c0; C MCD satisfy: c1 is the consistency coefficient; The identification module is used to determine that the deformation monitoring value is a normal value when the deformation residual of the concrete dam is less than or equal to the residual threshold and the robust distance of the deformation influencing factor is less than or equal to the robust distance threshold; otherwise, the deformation monitoring value is determined to be an outlier; wherein the deformation residual of the concrete dam is the absolute value of the difference between the deformation monitoring value and the deformation prediction value.
5. An electronic device, characterized in that: It comprises a processor and a memory coupled to the processor; the memory is used to store computer instructions, and when the electronic device is running, the processor executes the computer instructions stored in the memory, so that the electronic device executes the method as described in any one of claims 1 to 3.
6. A computer-readable storage medium, characterized in that: The method comprises computer program instructions, which, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 3.
7. A computer program product, characterized in that The method comprises computer program instructions, which, when executed on a computer, cause the computer to execute the method according to any one of claims 1 to 3.
